Forecasting & Predicting ABM

Forecasting and predicting ABMs with BINNs README file

System Requirements:

• Python 3
• High performance computing is recommended for ABM data generation and BINN training.
• ABM data generation was performed with 20 cores
• BINN training was performed using GPUs

Installation:

Following The Good Research Code Handbook, you can use pip to install the src package for this project. Once you have downloaded this code, you can install this package in the main directory directory by entering

pip install -e .

Running the ABM and training BINN models:

See the README.md files in scripts/Data_generation and scripts/BINN_training/ to see how to run the ABMs and train BINN models to pre-computed ABM data, respectively.

ABM forecasting and prediction:

ABM forecasting and prediction can be performed by running the jupyter notebooks located in scripts/Forecasting/ and scripts/predicting/, respectively.

Contact:

Please contact John Nardini at nardinij@tcnj.edu if you have any questions.